Research Activities #226, May 1999

influence physicians’ recommendations for cardiac catheterization, concludes a study supported by the Agency for Health Care Policy and Research (HS07315). In the study, the investigators found that blacks and women with chest pain had relative odds of referral for cardiac catheterization that were 60 percent of the odds for whites and men. This disparity in testing was greatest for black women, who had relative odds that were 40 percent of those for white men. Cardiac catheterization is the gold standard test for diagnosing coronary artery disease. During the procedure, dye is injected into the patient’s heart through a tube inserted into a major artery, usually in the groin. The path of the dye is then viewed through an x-ray image to show blood flow through the heart and locate arteries with blockages. Previous studies have been unable to assess whether physicians contribute to differences in access to care for blacks and women. This carefully controlled study used computer multimedia technology to address this question directly. Patient actors—two black men, two black women, two white men, and two white women—described their chest pain using the same scripts reporting identical clinical symptoms. They wore identical gowns, used similar hand gestures, and had the same insurance and professions. Researchers asked 720 primary care doctors at annual meetings of professional societies to interview a patient, review the patient’s medical data, assess his or her diagnosis, and recommend further diagnostic tests. Doctors were told that they were participating in a study of clinical decisionmaking. The study analysis was based on a research model assessing recommendations for cardiac catheterization incorporating six experimental factors and physicians’ assessments of the probability of coronary disease for the patient they interviewed. In further analyses, the investigators determined that the results were unchanged when information on physicians’ perceptions of patient personalities and physician characteristics were included in the model. Biases found in this study may reflect subconscious perceptions of physicians about Research Activities